Parallelize 'SingleCellExperiment' functions

Henrik Bengtsson

+ =

The futurize package allows you to easily turn sequential code into parallel code by piping the sequential code to the futurize() function. Easy!

TL;DR

library(futurize)
plan(multisession)
library(SingleCellExperiment)
library(scuttle)

result <- applySCE(sce, perCellQCMetrics) |> futurize()

Introduction

This vignette demonstrates how to use this approach to parallelize the SingleCellExperiment functions.

The SingleCellExperiment Bioconductor package defines the SingleCellExperiment class for storing single-cell genomics data, including alternative experiments (e.g. spike-in transcripts, antibody tags). The applySCE() function applies a given function to the main experiment and each alternative experiment, passing additional arguments such as BPPARAM via ... to enable parallelization of the applied function.

Example: Computing per-cell QC metrics in parallel

The applySCE() function applies a function across the main experiment and its alternative experiments:

library(SingleCellExperiment)
library(scuttle)

# Simulate data
set.seed(42)
n_genes <- 200L
n_cells <- 100L
counts <- matrix(
  rpois(n_genes * n_cells, lambda = 10),
  nrow = n_genes,
  ncol = n_cells,
  dimnames = list(
    paste0("gene", seq_len(n_genes)),
    paste0("cell", seq_len(n_cells))
  )
)

sce <- SingleCellExperiment(
  assays = list(counts = counts)
)

# Add an alternative experiment (e.g. spike-ins)
spike_counts <- matrix(
  rpois(10L * n_cells, lambda = 5),
  nrow = 10L,
  ncol = n_cells
)
rownames(spike_counts) <- paste0("spike", seq_len(10L))
colnames(spike_counts) <- paste0("cell", seq_len(n_cells))

altExp(sce, "spikes") <- SingleCellExperiment(
  assays = list(counts = spike_counts)
)

result <- applySCE(sce, perCellQCMetrics)

Here applySCE() runs perCellQCMetrics() sequentially on each experiment, but we can easily make it run in parallel by piping to futurize():

library(futurize)

result <- applySCE(sce, perCellQCMetrics) |> futurize()

This will distribute the work across the available parallel workers, given that we have set up parallel workers, e.g.

plan(multisession)

The built-in multisession backend parallelizes on your local computer and works on all operating systems. There are other parallel backends to choose from, including alternatives to parallelize locally as well as distributed across remote machines, e.g.

plan(future.mirai::mirai_multisession)

and

plan(future.batchtools::batchtools_slurm)

Supported Functions

The following SingleCellExperiment functions are supported by futurize():

• applySCE()